a2ui-cloud-run-deployer

name	A2UI Cloud Run Deployer
description	Deploys an A2UI ADK agent cleanly to Google Cloud Run in namespaced A2A-compatibility mode.

A2UI Cloud Run Deployer Skill

This skill automates clean container packaging, compilation (via Cloud Build), and namespaced A2A deployments of A2UI agents to Google Cloud Run.

Instructions

When this skill is triggered, you must execute the following steps in sequence:

Step 1: Input Collection & Verification

Ask the user for the following deployment variables:

• GCP_PROJECT_ID: The Google Cloud Project ID.
• GCP_REGION: The deployment region (defaults to us-central1 if empty).
• AGENT_DIR: The path to the target agent folder in the workspace.
• SERVICE_NAME: The Cloud Run service name (defaults to the agent folder name).

Step 2: Clean Staging & Deployment Execution

Run the helper deployment script located in this skill's scripts directory:

python3 .agents/skills/a2ui_cloud_run_deployer/scripts/deploy_a2ui.py \
  --project <GCP_PROJECT_ID> \
  --region <GCP_REGION> \
  --agent_dir <AGENT_DIR> \
  --service_name <SERVICE_NAME>

Step 3: Verify the Deployment

Once the deployment script outputs success:

1. Retrieve Agent Card: Query the Agent Card URL via curl to verify availability:

   curl -s https://<service-url>/a2a/<agent_name>/.well-known/agent-card.json
   

2. Verify Agent Execution: Send a basic JSON-RPC 2.0 test message to ensure the agent is responsive:

   curl -X POST \
     -H "Content-Type: application/json" \
     -d '{"jsonrpc": "2.0", "method": "message/send", "params": {"message": {"role": "user", "parts": [{"text": "Hello"}], "messageId": "test-msg-id", "contextId": "test-session-id"}}, "id": 1}' \
     https://<service-url>/a2a/<agent_name>
   

Step 4: Display Output Endpoints

Print a clean summary displaying the final endpoints:

• Service URL: https://<service-url>
• Namespaced A2A RPC URL: https://<service-url>/a2a/<agent_name>
• Agent Card Endpoint: https://<service-url>/a2a/<agent_name>/.well-known/agent-card.json